Following a successful pilot phase, Quebec’s Ministry of Education (MEQ), in collaboration with GRICS, has begun deploying a predictive analytics tool designed to detect risk factors for student dropout from Grade 6 through Secondary V. Managed by the Centre d’expertise de l’intelligence en éducation (CEIAE), the project is already being refined based on feedback from early adopters in the field.
The tool conducts a multidimensional analysis of existing student data—such as grades and attendance records—to flag those most at risk. Dashboards give educators and school leaders greater visibility to act proactively, aligning supports with student needs.
That’s the key message from Natalia Cruz Currea, MEQ’s coordinator for data governance and AI support in education. She emphasized that educators remain responsible for interpreting results and making decisions, with human validation built into the tool’s process at several stages.
To date, 60 School Service Centres (CSS) have expressed interest, with 22 having fully implemented the tool and 27 others in active rollout. Each deployment follows a nine-week sequence, beginning with quick access to dashboards in the first month and followed by training and personalized support.
Feedback has been positive. According to Cruz Currea, the tool is seen as both intuitive and relevant, especially for spotting “silent dropouts”—students who may not exhibit obvious signs of disengagement. Each flagged case also comes with research-backed suggestions for intervention, helping school teams respond effectively.
Data privacy remains a top priority. Hosted securely on Microsoft Azure servers in Canada, the tool meets rigorous cybersecurity standards. Access is limited to authorized personnel: CSS directors can view their system-wide data, while school leaders see only their own school’s.
Importantly, all data used by the tool was already being collected by schools, and no explicit consent is needed under Quebec’s Law 25, as the data remains within its original educational purpose. Parents are informed about how the tool works.
In collaboration with research leaders such as Mila, Obvia, and CTREQ, the MEQ is working to enhance both the ethical and practical dimensions of the AI system. The next phase aims to incorporate more data sources, including school transport records and observations logged via GRICS’ SOI (Suivi des observations et des interventions) platform.
The tool currently targets a 70% accuracy rate, but pilot testing in the Outaouais region yielded a remarkable 92%—a promising result that now needs validation across more CSSs and longer timelines.
Though currently focused on elementary and secondary levels, the tool may eventually extend to adult general education (AGE) and vocational training (VT), which would require model adjustments due to different student profiles.
The long-term goal is to expand beyond individual interventions and use AI-driven insights to shape broader school- and system-level strategies. As Cruz Currea explains, “By identifying system-wide trends, leaders can deploy more effective, large-scale solutions to prevent school dropout across Quebec.”
📌 This story stems from the inaugural Performance et valorisation des données en éducation (Performance and Data Use in Education) conference held on March 19, 2025, at the Quebec City Convention Centre, hosted by Le Point en Santé | Éducation and partners.
If you are subscribed, log in to access exclusive member content!